Video Steganalysis Algorithm Based On Bispectrum

In recent years, along with the rapid development of digital multimediatechnology and network technology, digital products have brought much convenienceto people’s daily work and life. Meanwhile, all kinds of information security problemsemerge in endlessly. In order to improve this situation, since the1990s, informationhiding has begun and developed rapidly. Using video as the cover, video informationhiding refers to embedding undetectable secret information into the cover toaccomplish concealed communication. As the attacking technology, videosteganalysis technology aims at revealing the existence of secret information, whichhas important significance for the development and improvement of informationsecurity technology.Among the video information hiding technologies, video steganography based onspread spectrum is an important representative, which can resist compression attackseffectively and has strong robustness and security. Aiming at detecting thistechnology, a video steganalysis method based on bispectrum is proposed in thispaper, for which the main purpose is to detect the existence of secret information andestimate the embedded strength. Under the full use of strong time correlation of thevideo sequence, this paper utilizes the difference model as prediction model. Throughestimating the bispectrum of the difference model, the frequency value of theminimum point in the center line is used to estimate the strength of the embeddedinformation. Finally, in order to validate the feasibility and robustness of the proposedalgorithm, Matlab is used to do the simulation experiment. Experimental resultsdemonstrate the effectiveness of this approach.